2505
J. Am. Chem. SOC.1993, 115, 2505-2507
for carbons attached to, respectively, one and two nitrogen atoms." The IH signal (6, 7.72) for the hydrogen attached to the former carbon (6, 128.0) showed long-range coupling (J = 1.O Hz) to an anomeric proton (6, 4.89, H-1'). These data suggested that the base unit is either pyrrolo[2,3-d]pyrimidine (Le., 7-deazaadenine) or pyrrolo[ 3,2-d]pyrimidine (Le., 9-deazaadenine). The I3C signals due to the aromatic carbons of 1 resemble those reported for 9-deazaadenine derivatives'* rather than 7-deazaadenine derivatives," although no pyrrolo[ 3,2-d]pyrimidine derivative has been reported from natural source^.^ UV spectra of 1, especially the shifts of absorption maxima in acidic solution, were also more like those of 9-deazaadenineI4 than those of tubercidin (7-deazaadeno~ine).'~Accordingly, the base unit in 1 is most likely 9-deazaadenine. The I3C signals assigned to the hexose unit of 1 closely resembled those of methyl a-D-glucopyranoside,I6 suggesting that 1 is the a-D-glucopyranoside of 9-deazaadenosine. 5'-ff-DGlucopyranosides of tubercidin and toyocamycin (3 and 4, respectively, Scheme 11) have been isolated from cyanobacteria." ' H and 13CN M R data for the sugar units of 1 were very similar to those for 3 and 4, except for the signals due to the C-1' position. Moreover, enzymatic deglycosidation of 1 with a-D-glucosidase gave &glucose and 2, which was isolated as the minor component (1 1% of 1) from the same cyanobacterium. Compound 2, [aIBD-28.4O (c 0.016, HzO),showed a molecular ion peak at m / z 267.1090 (CllHISN404, M H, A +0.3 mDa) by HRFABMS. FABMS/CID/MS of 2 gave the same fragment ion peaks at m / z 177, 163, and 147 observed for 1 (Scheme I). The IH N M R spectrum of 2 showed the signals ascribable to a ribose unit and two aromatic proton signals.I* From the results above, the structure of 2 can be assigned as 9-deazaadenosine, which has been synthesized by Lim and Klein as a cytotoxic C-nucleoside isostere of adenosine.Ig The direct comparison of 2 with a synthetic sample of 9-deazaadenosineZ0 by HPLC, TLC, and UV spectra confirmed that 2 was identical to synthetic 9-deazaadenosineP ' H N M R data for natural 2 hydrochloride were also identical with those for synthetic 2 hydrochloride.21 Consequently, the structure of 1 was assigned as the 9-deazaadenosine 5'-a-~-glucopyranoside, as shown in Scheme I. Compounds 1 and 2 are pyrrolo[3,2-d]pyrimidinederivatives which have not been reported previously as biosynthetic p r o d ~ c t si.e., ,~ from natural sources. Their biosynthesis will be of considerable interest. The I C d of 1 and 2 vs L1210 murine leukemia cells were 0.01 and 0.002 pg/mL, respectively. These compounds also showed
+
( 1 1) Breitmaier, E.; Voelter, W. Carbon-I3 N M R Spectroscopy; VCH: Weinheim, Germany, 1987; p 289. (12) Buchanan, J. G.; Craven, D. A,; Wightman, R. H.; Harnden, M. R. J . Chem. SOC.,Perkin Trans. I 1991, 195-202. (13) Chenon, M.-T.; Pugmire, R. J.; Grant, D. M.; Panzica, R. P.; Townsend, L. B. J. Am. Chem. SOC.1975, 97, 4627-4636. (14) Imai, K. Chem. Pharm. Bull. 1964, 12, 1030-1042. (15) Anzai, K.; Nakamura, G.; Suzuki, S. J . Antibiof., Ser. A 1957, 10, 20 1-204. (16) Bock, K.; Pedersen, C. In Advances in Carbohydrate Chemistry and Biochemistry; Tipson, R. S.;Horton, D., Eds.; Academic: New York, 1983; Vol. 41, pp 27-66. (17) Stewart, J. B.; Bornemann, V.; Chen, J. L.; Moore, R. E.; Caplan, F. R.; Karuso, H.; Larsen, L. K.;Patterson, G. M. L. J. Antibiof. 1988, 41, 1048-1056. The assignments of IH and 13C signals for the C-3" and 4" positions must be interchanged. (18) 'H NMR data (500 MHz, 18 "C) for 2 in DMSO-d6 (2.49 ppm): 6 11.95(1 H , b r s , N H ) , 8 . 1 7 ( 1 H , s , H - 2 ) , 7 . 7 2 ( 2 H , b r s , N H 2 ) , 7 . 5 8 ( 1 H, S, H-6), 4.85 ( 1 H, d, J = 3.1 Hz, OH), 4.77 ( I H, d, J = 7.4 Hz, H-l'), 4.22 ( 1 H,dd, J = 7.4. 5.1 Hz, H-2'), 4.14 (1 H, d, J = 3.9 Hz,OH), 4.00 (1 H, dd,J=5.1,2.8Hz,H-3'),3.86(1 H , d d d , J = 3 . 1 , 3 . 1 , 2 . 8 H ~ , H - 4 ' ) , 3 . 6 0 (1 H, dd, J = 12.0, 3.1 Hz, H-5'), 3.51 ( 1 H, dd, J 12.0, 3.1 Hz, H-5'); assigned by single-frequency decoupling experiments. (19) Lim, M.-I.;Klein, R. S . Tetrahedron Lett. 1981, 22, 25-28. (20) The synthetic sample of 9deazaadencsine hydrochloride was provided by Dr. Robert S. Klein, Montefiore Medical Center. (21) 'H NMR data (500 MHz, 18 "C) for 2 hydrochloride in DMSO-d6 (2.49 ppm): 6 12.80 (1 H, s, NH), 9.03 and 8.99 (each 1 H, s, NH2), 8.50 ( 1 H, S. H-2), 7.86 ( I H, d, J = 1.0 Hz, H-6), 4.86 (1 H, d, J = 7.0 Hz, H-1'), 3.97 ( 1 H, dd, J = 7.0, 5.1 Hz, H-3'). 3.94 ( 1 H, dd, J 5.1, 3.1 Hz, H-2'), 3.87 ( 1 H, dt, J = 3.2, 3.1 Hz, H-4'), 3.62 (2 H, d, J = 3.2 Hz, Hz-5').
lethal toxicity to the aquatic invertebrate Ceriodaphnia dubia; the LCS$ for acute (48 h) and chronic (7 day) toxicities were, respectively, 0.5 and 0.3 Mg/mL for 1 and 0.3 and 0.1 pg/mL for 2. Acknowledgment. This study was supported by a grant from the National Institute of Allergy and Infectious Diseases (AI 04769, to KLR) and a subcontract from the same grant to W. W.C. We thank Dr. Robert S. Klein of Montefiore Medical Center for the synthetic sample of 9deazaadenosine hydrochloride and for useful information and Dr. John Gilbert of Dartmouth College for the culture of A . affinis strain VS-1. Supplementary Material Available: ' H NMR, FABMS/ CID/MS, and UV spectra of 1 and 2 I3C NMR, 'H-IH COSY, 'H-I3C COSY spectra of 1; and 'H NMR spectra of natural and synthetic 2 hydrochloride (12 pages). Ordering information is given on any current masthead page.
[c& iSO-cfi7+]
and [c&+C&] Ion-Molecule Complexes: Theoretical Calculations D. Berthomieu,**' V. Brenner,t G. Ohanessian,l J. P. Denhez,$ P. Millii,; and H. E. Audiers Loboratoire de Chimie Organique Structurale CNRS URA 455. Universite P. et M.Curie, 4 place Jussieu 75005 Paris, France CEA-CE Saclay, DSM/DRECAM/SPAM 91 191 GIF SUR Yvette, France DCMR, Ecole Polytechnique, CNRS URA 1307 91 128 Palaiseau Cedex, France Received July 2, 1992
While arenium ions are known to be intermediates in electrophilic aromatic substitution reactions,' the existence of ?r-complexes between an aromatic group and a cation remains elusive. In the gas phase the intermediacy of *-complexes has often been postulated in the unimolecular fragmentation of aromatic cations,2-8 but there is as yet no irrefutable evidence for their existence. The existence of ion-molecule complexes [C6H7+alkene] has been proposed from experimental results.8 In this work we have, for the first time, calculated the energy and structure of *-complexes and ion-molecule complexes involving the benzenium ion and an alkene. This kind of system is a good example of the use of molecular orbital calculations in order to calculate the energy and to study the structure of ion-neutral complexes. This has been recently r e ~ i e w e d . ~ We have chosen to focus on one model: the complex intermediates presumed to be involved in the unimolecular reaction of metastablelo protonated isopropylbenzene. It has been pre'Universit€ P. et M. Curie. CEA-CE Saclay. Ecole Polytechnique. ( I ) Norman, R. 0. G.; Taylor, R. Electrophilic Substitution in Benzenoid Compounds; American Elsevier: New York, 1965. (2) For a review, see: Kuck, D. Mass Specfrom.Reu. 1990. 9, 583-630, and references cited therein. (3) BBther, W.; Grutzmacher, H. F. I n f . J. Mass Spectrom. Ion Proc. 1985, 64, 193-212. (4) Kuck, D.; Matthias, C. J. Am. Chem. SOC.1992, 114, 1901-1903. ( 5 ) Herman, J . A.; Harrison, A. G. Org. Mass Specfrom. 1981, 16, 423-427. (6) Speranza, M. Specfrosc.Int. J. 1987, 5 , 1 and references cited therein. (7) Holman, R. W.; Gross, M. L. J. Am. Chem. SOC. 1989, 1 1 1 , 3560-3565. (8) Berthomieu, D.; Audier, H.-E.; Denhez, J.-P.; Monteiro, C.; Mourgues, P. Org. Mass Spectrom. 1991, 26, 271-275. (9) Longevialle, P. Mass Spectrom. Reu. 1992, 1 1 , 157-192.
0002-786319311515-2505$04.00/0 0 1993 American Chemical Society
Communications to the Editor
2506 J. Am. Chem. Soc.. Vol. 115, No. 6,1993
(a)
(b)
Figure 1. Structures for (a) the [C6H6 iso-ClH7+] complex and (b) the [C6H7+ ClH6] complex. In (a), the d value is 3.1 1 A, with the a b initio calculation and 2.95 A with the direct calculation. In (b), the d value is 2.51 A with the a b initio calculation and 2.02 A with the direct calculation.
viously shown8 that protonated n-propylbenzene isomerizes irreversibly to protonated isopropylbenzene prior to dissociation. Both cations lead to C6H7+(87%) and iso-C3H7+ (1 3%) product ions (eq 1). Dissociations are preceded by an incomplete but specific
Figure 2. Potential energy surface for the unimolccular decomposition of metastable protonated isopropylbenzene. Bold characters correspond to AHfo in kJ/mol.'* The numbers in parentheses correspond to stabilization energies calculated by the semiempirical method.
H-exchange between the primary and aromatic hydrogen atoms. These data can be explained only if a [C6H7+C3H6]complex, hereafter called 8, is postulated (eq lb). The [C6H6 iso-C3H7+] complex, hereafter called r , may be the precursor of iso-C3H7+ (eq la). In this model the H-exchange can occur by reversible isomerization between protonated isopropylbenzene and the 8complex, between r - and @-complexes, or both. First, a b initio calculations" were carried out a t the 631G**//3-21G level and led to the existence of local minima corresponding to both @- and r-complexes. All optimized structures were characterized as true minima through vibrational energy analysis at the 3-21G level. After correction for zero-point vibrational energy, the well depths are 39 kJ/mol for the r complex and 13 kJ/mol for the &complex, relative to final products (values before correction are 46 and 20 kJ/mol, respectively). However, ab initio calculations are intractable for a complete exploration of the potential energy surface (PES) and remain approximate since electron correlation calculations are not feasible at present. In such complexes, the single most important contribution to electron correlation is the dispersion energy. We thus performed direct calculations of the interaction energy calculated as the sum of terms: electrostatic, polarization, dispersion, short-range repulsion, and exchange dispersion energies, according to the method developed by C l a ~ e r i e . l ~We - ~ ~then determinated the most significant minima on the PES on the basis of a simulated annealing method. 5-'
From the simulated annealing calculations, only one stable form has been found for the [C6H6iw-CJH7+]complex. This structure is almost the same as the ab initio optimized structure. However, due to dispersion forces, the structure is a little more compact. The interaction energy is 67 kJ/mol, and the structure of this complex corresponds to a *-complex. In fact, the isopropyl cation is centered on the benzene, with its three carbon atoms inclined with respect to the plane of the six carbon atoms of the benzene (Figure 1). The dihedral angle between these two planes is 40°. The distances between the center of gravity of the benzene and the secondary carbon are 2.95 and 2.39 A, respectively, for the hydrogen atom linked to this carbon. The calculations for the 8-complex [C6H7+C3H6]show many minima on the PES. For the most stable 6-complex the interaction energy is 32 kJ/mol (Figure 1). As in the case of the r-complexes, the geometry is very similar to ab initio values with shorter distances between the two components. One hydrogen bonded to an sp3 carbon is in a favorable sition for further transfer to the propene molecule (d = 2.02 Figure 1). This distance is a little smaller than with the ab initio calculation. The important number of local minima for the 6-complexes obtained from the simulated annealing study corresponds to the interaction of the propene with almost every hydrogen atom of the protonated benzene. There is no minimum with the propene molecule above the protonated benzene. For all these complexes, intermolecular distances smaller than 2.3 A are observed between the two components. From these calculations we can now consider the potential energy profile of the fragmentation reactions (Figure 2). The stabilization energy of the *-complex is 67 kJ/mol, and it is 32 kJ/mol for the &complex, both relative to final products.l* The *-complex lies below the most stable final state, corresponding to formation of C6H7+. This result leads us to conclude that the intermediacy of T - and ,!%complexesis certainly a pathway leading to H-exchange and therefore to the formation of iso-C3H7+and C6H7+from low-energy protonated isopropyl benzene ions. Metastable protonated tert-butylbenzene yields tert-C4H9+ without H-exchange. Simulated annealing calculations show that T - and @-complexescorrespond to stable structures. However, the &complex lies higher in energy than the final state (tert-C4H9+
(IO) Cooks, R. G.; Beynon, J. 13.; Caprioli, R. M.; Lester, G . R. Metastable Ions; American Elsevier: New York, 1973. (1 1) Hehre, W. J.; Radom, L.;Schleyer, P. v. R.; Pople, J. A. Ab Initio Molecular Orbital Theory; John Wiley & Sons: New York, 1986. (12) Claverie, P. In Intermolecular Interactions; Pullman, 8.. Ed.; John Wiley & Sons: New York, 1978; Chapter 3. (13) Brenner, V.;Martrenchard, S.; Milli6, P.; Jouvet, C. Chem. Phys. 1992, 162, 303-320. (14) The total procedure is the following. First, a configuration space exploration by use of the Metropolis algorithm allows the selection of some energetically favored configurations. This selection is made at each fictitious temperature T which is slowly decreasing along this process. Second, these configurations are used as starting points for a minimization procedure by a standard local minimization method. Thus, one generally obtains some local minima. Third, each local minimum is probed by a scrutiny of the Hessian eigenvalues.
(15) Press, W. H.; Flannery, B. P.; Teukolsky, S. A,; Vctterling, W. T. Numerical Recipes, The Art of Scientific Computing; University Press: Cambridge, 1986; Chapter 10. (16) Kirkpatrick, P. J. Stat. Phys. 1984, 34, 975. (17) The electronic distribution obtained by the SCF calculation of each molecule or ion is superseded by a monopole, dipole, and quadrupole centered on each atom and at the middle of each bond so that the true SCF electrostatic potential is correctly fitted. The polarization energy is calculated by using experimental atomic and bond polarizabilities. The repulsion-dispersion energy is calculated as the sum of atom-atom repulsion-dispersion energy terms. Thus, the calculation of the interaction energy and its derivatives is achieved by using analytical formulae. In this way, simulated annealing and quasi-Newtonian local minimization calculations are easily tractable, even for such complex systems. (18) Lias, S.G . ; Liebman, J. F.: Levin, R. D. J. Phys. Chem. Re/. Data 1984, 13, 695-808.
-
CSH&H~CH~CH~* AHf' 733
P AHf* are in kJlmol
075
'
J. Am. Chem. SOC.1993, 115, 2507-2508
2507
plus benzene), which explains why H-exchange does not occur. The interaction energy for the more stable *-complex structure is 56 kJ/mol. The stability of *-complexes is a general phenomenon. In a further publication we will give other theoretical and experimental evidences which confirm their existence in related cases.
Crystal Structure and Reactivity of a Pentacoordinate 1,2-Oxasta~etanide:An Intermediate of the Tin-Peterson Reaction Takayuki Kawashima,* Naoshi Iwama, and Renji Okazaki; Department of Chemistry, Faculty of Science The University of Tokyo 7-3- 1 Hongo, Bunkyo- ku, Tokyo 1 1 3, Japan Received December 4 , 1992 The Peterson reaction has been widely utilized for olefin synthesis as a silicon analog of the Wittig or Horner-Emmons reaction, providing the method for selective synthesis of (E)- or (a-isomer from a single diastereomer of (8-hydroxyalky1)silanes.' The reactions using homologs such as @-hydroxy germanes, stannanes, and plumbanes are well-known to give the corresponding olefins under acidic and neutral (or basic) conditions.le** Very recently, we succeeded in the synthesis of pentacoordinate 1,2-oxaphosphetane l3and 1,2-oxasiletaNde 2 bearing the Martin ligand, intermediates of the Wittig and the Peterson reactions, respectively.
Figure 1. ORTEP drawing of Sa (omitting CH2C12and H 2 0 ) . Selected bond lengths (A) and bond angles (deg): S n ( l ) - O ( l ) , 2.401(5); Sn( 1 ) 4 ( 7 ) , 2.140(7); S n ( l ) C ( l 3 ) , 2.1 36(8); S n ( l ) C ( 1 9 ) , 2.188(8); S n ( l ) C ( 2 5 ) , 2.200(7); O(l)-C(26), 1.370(8); C ( 2 S ) C ( 2 6 ) , lSS(1); O(1)-Sn( 1 ) C ( 19), 165.1(2); C(7)-Sn( I)
